设备故障诊断中的特征提取

Feature extraction of equipment fault diagnosis

以某航天器姿控系统为研究对象,对故障的原始特征进行特征提取和选择,以使所建立的故障标准模式由少数几个新特征给予有效的表达,较好地实现了主要故障模式的分离,为提高设备故障诊断能力和故障定位能力奠定了基础。同时,通过考察权系数,可判断出主分量中各个参数的贡献率,这将有利于指导设计阶段的测点选择。

Taking the attitude control system of the spacecraft as the research object, the original feature of the fault was extracted and chosen, so the normal fault mode was effectively expressed by several new features, thus separation of the main fault mode was developed, which establishes the base of increasing the ability of equipment fault diagnosis and orientation. At the same time, the contribution ratio of every parameter in main component could be judged through the weight factor, which is favorable for measuring point choice in design stage.